Open AccessПрименение искусственной нейронной сети в экспериментах по ядерному спиновому эху
Author(s) -
С.Н. Полулях,
А.И. Горбованов
Publication year - 2019
Publication title -
pisʹma v žurnal tehničeskoj fiziki
Language(s) - English
Resource type - Journals
eISSN - 1726-7471
pISSN - 0320-0116
DOI - 10.21883/pjtf.2019.12.47911.17736
Subject(s) - echo (communications protocol) , amplitude , artificial neural network , noise (video) , gaussian noise , acoustics , computer science , physics , speech recognition , artificial intelligence , optics , computer network , image (mathematics)
The possibility of artificial neural network application to detect nuclear spin echo signals under conditions when the echo amplitude is comparable to the amplitude of the noise is demonstrated. Data obtained by superimposing the model echo signals of a Gaussian form on experimentally recorded noise signals is proposed to use for training the neural network.